How do you explain why you chose a specific statistical model or how you handled missing data in a past project?